Mechanisms of SNi Reactions. The Decomposition of Aralkyl Thiocarbonates' John L. Kice, Richard A. Bartsch, Mary A. Dankleff, and Sandra L. Schwartz
Contribution f r o m the Department of Chemistry, Oregon State University, Corvallis, Oregon. Received January 8, I965 Thiocarbonates ( R O C O S R ' , R = aralkyl) decompose to carbon dioxide and the corresponding sulfides ( R S R ' ) on being heated at elevated temperatures in inert solvents. The decomposition is clearly analogous to the decomposition of alkyl chlorocarbonates to alkyl chlorides and carbon dioxide,2t and study of the dependence of the rate of decomposition on both solvent and the nature of the aralkyl group shows that, like the chlorocarbonate decomposition, the reaction involves rate-determining formation of the aralkyl carbonium ion b y heterolysis of the R - 0 bond. The rate of decomposition of a series of benzhydryl thiocarbonates, PhZCHOCOSR', varies markedly with R ' . Of particular interest is the fact that the, compound where R ' equals phenyl decomposes more rapidly than would be expected f r o m the decomposition rates of a series of S-alkyl thiocarbonates. This result seems most readily explained b y a mechanism in which both the R-0 and C-S bonds are broken in the rate-determining step. The possible implications f o r the mechanism of the chlorocarbonate decomposition and other Sivi reactions are discussed. The decomposition of alkyl chlorocarbonates (eq. 1) is one of the classic examples of an SNi reaction. The variation of the rate of decomposition with the nature R-O-C-CI
ii
+R-CI
+ COz
(1)
of the alkyl group2, and with solvent ionizing power2 make it clear that the rate-determining step of the
bond is synchronous with (eq. 2a), or alternatively occurs subsequent to (eq. 2b), the fission of the R-0 bond.
rate
R-O-CCI
a
+[R+ -OzC-Cl]
COz
-+
+ [R+ CI-]
-+
R-CI
(2b)
detg.
One way to obtain information pertinent to this point is to find a compound of the structure ROCOZRi which undergoes an analogous decomposition (ROCOZR' RZR' CO,) and to investigate the influence of suitable changes in R ' on its rate of decomposition. We have found that thiocarbonates of the structure ROCOSR ' undergo thermal decomposition at elevated temperatures in a reaction (eq. 4) which is not only formally analogous to the chlorocarbonate decomposition but also shows a similar dependence on solvent ionizing power and structure of the R group. Accordingly, the thiocarbonate decomposition seems ideally suited for studies of the type outlined above. The present paper describes the results of studies of the variation of thiocarbonate decomposition rate with variation in R ' and discusses the implications of those results as regards the general question of the timing of the various bond scissions involved in a typical SNi reaction. --f
+
Table I. Products of the Decomposition of Aralkyl Thiocarbonates
Carbon dioxide RSR' R'SSR' R-R ROH = Mole/mole of thiocarbonate.
1.01 0.83 0.02 0.07
0.00
0.94 0.67 b b 0.07
* None detected, but a small amount could possibly have been present.
reaction involves heterolysis of the R-0 bond and formation of the alkyl carbonium ion R+. It has not been known, however, whether breaking of the c - c l (1) (a) This research was supported by the National Science Foundation under grants G-19503 and GP-1975. (b) Part of this work has aooeared in oreliminarv form: J. L. Kice and R . A . Bartsch. Tetrahedron Lkiters, 169j (1963). (2) K . L. Olivier and W. G . Young, J . A m . Chem. SOC.,81, 5811 (1959). (3) K.B.Wiberg and T. M. Shryne, ibid., 77, 2774 (1955).
Results Synthesis of Thiocarbonates, The various thiocarbonates were all synthesized by reaction of the amrotxiate alcohol with the Drot3er chlorothiolformate .. (eq. $. Data on their ph;sical constants, etc., are given in Table v.
~
1734
1 .OO 0.78 0.02 0.02 0.00
0.95 0.87 b b 0.00
0.95 0.83 b b 0.00
Journal of the American Chemical Society
1 87:8
April 20, 1965
R-OH
+ R'SX-C1 II
0
pyridine
R-OC-SR'
I1
0
(3)
Table 11. Kinetics of the Decomposition of Aralkyl Thiocarbonates
--
Thiocarbonate, ROCOSR' R R'
7-
Solvent
C&Br
134 145 145
CeHsCN
155 166
CsHsCN
166
CeHsCN
166
CeHsCN
166
CsHsCN
166
CeHbCN CeHsCN CsHbCN
123 131 139 186 146
CeHsCN
166
CeHjCN
kl
[ROCOSR'Io, M
Temp., "C.
0.081 0.066 0,088 0,048 0.088 0.072 0.053 0.075 0.061 0.072 0.050 0.145 0.086 0.10 0.13 0.10 0.10 0.10 0.072 0.087
1.8 4.6 0.18 0.19 0.45 2.4 2.4 1.1 1.2 0.66 0.66 4.7 4.7 5.5 7.4 0.27 0.60 1.2 0.0042 2.4
0.068
0.57 0.65
0,055
Decomposition of Aralkyl Thiocarbonates. On being heated at elevated temperatures in inert solvents aralkyl thiocarbonates undergo decomposition with essentially quantitative evolution of carbon dioxide (eq. 4). R-04-SR'
a
+ R-SR'
+ COz
(4)
The appropriate organic sulfide can be isolated as the other major reaction product in yields ranging from 6 7 - 8 7 z (Table I). Small amounts of other products were isolated in certain cases. These may result not only from side reactions involving the thiocarbonate but also from some breakdown of the aralkyl sulfide at the temperatures involved. In any event, Table I shows that in all five cases where a detailed product study has been made the principal course of the decomposition is that shown in eq. 4. We assume this is also true for the other aralkyl thiocarbonates whose decompositions have been investigated kinetically in the present work. Kinetics of the Decomposition. For all the aralkyl thiocarbonates except benzhydryl S-carbomethoxymethyl thiocarbonate the kinetics of the decomposition can be followed by observing the disappearance of the infrared absorption band due to the carbonyl group of the thiocarbonate. Depending on the thiocarbonate, this occurs at 1705 to 1720 cm.-'. When followed in this way the decompositions were found to exhibit good first-order kinetics over at least 75 % reaction. With some of the slower runs there was a tendency for the rate to accelerate somewhat after this point, apparently due to catalysis of the decomposition by substances resulting from the breakdown of a small amount of the normal decomposition products. In these cases the rate constant was determined from the initial linear portion of the log c vs. time plot. Some typical runs are shown in Figure 1.
x 104, sec.-l
In the case of the S-carbomethoxymethyl ester an alternate procedure, employing measurement of the rate of evolution of carbon dioxide, was used to follow the reaction. A run using this procedure with benzhydryl S-phenyl thiocarbonate gave a rate constant identical with that obtained by the infrared method. The various kinetic results are summarized in Table 11.
$0
loo
1%
mo
Tima ( d n )
Figure 1. First-order plots of kinetic data for some representative thiocarbonate decompositions : 0, benzhydryl S-phenyl thiocarbonate, 0.060 M , 145' ; 8 , benzhydryl S-benzyl thiocarbonate, 0.072 M , 166"; 0 , benzhydryl S-methyl thiocarbonate, 0.061 M , 166"; and a, benzhydryl S-ethyl thiocarbonate, 0.05 M , 166".
Dependence of Rate on Solvent. The rate of decomposition of benzhydryl S-phenyl thiocarbonate was determined in bromobenzene and benzonitrile at 145O . As is evident from Table 11, the rate is about 25 times greater in benzonitrile. This dependence of rate on solvent ionizing power seems comparable to that observed by Olivier and Young2 for the chlorocarbonate decomposition. Variation of Rate with Aralkyl Group. The variation of the rate of decomposition with nature of the aralkyl group was investigated for a series of aralkyl S-phenyl
Kice, Bartsch, Danklef, Schwartz
Decomposition of Aralkyl Thiocarbonates
1735
Table 111. Dependence of Rate on Structure of Aralkyl Group Relative rates
Decompn. of Decompn. of ROCOSCSH, ROCOCI in Solvolysis of in C&CN RC1 in HOAc dioxane at 250a at 90°b at 186”
Aralkyl group, R C6HjCHzC6HjCH-
(1 .O)
(1 .O) 370
...
(1 .O) 450
~
CH3 C6HjCHp-CICeHaCH-
2.6 1.2
I
x x
Table IV. Rate of Decomposition of (C6H6)CHOCOSR’in Benzonitrile at 166’
7
104 1 . 5 x 104 104~ 0.51 x 104
R’ CH3CHzCH3Cf,H,CHz(CBH~ZCHCICHzCHZCHgOOCCHzC6H5-
kl X IO4, set.-'
Relative rate
0.66 1.1 2.4 4.7 6.3 11 25
0.60 (1 .O) 2.2 4.3 5.7 10 23
...
...
C6Hs a Data for benzyl and 1-phenylethyl chlorides from A. Streitwieser, Chem. Rev., 56,616 (1956), that for benzhydryl chloride from S. Winstein, A. H. Fainberg, and E. Grunwald, J . Am. Chem. SOC.,79, 4146 (1957), and that for p-chlorobenzhydryl chloride from S. Winstein, et a/., ibid., 82, 1010 (1960). Reference 3. Calcd. from data at 146” assuming rate shows same temperature coefficient as for benzhydryl ester.
thiocarbonates, ROCOSPh. Table I11 shows their relative rates of decomposition at 186”, and for comparison, both the relative rates of acetolysis of a series of aralkyl chlorides and the relative rates of decomposition of two aralkyl chlorocarbonates. From these data it appears that the variation of the rate with the structure of the aralkyl group in the thiocarbonate
Figure 2 shows a plot of log k , vs. u* for R’4 for the thiocarbonates in Table IV. Despite the considerable variation in the steric requirements of R ’ the data for the six alkyl or aralkyl esters are quite well correlated. Judged by the behavior of these esters the phenyl compound decomposes more rapidly than would be expected, and the point for it deviates from the log kl-u* correlation line by considerably more than the points for any of the other esters. Discussion Analogous to eq. 2a and 2b, one can envisage two possible mechanistic extremes for the thiocarbonate decomposition. In one of these (eq. 5b) the ratedetermining step involves heterolysis of only the R-0 bond. In the other (eq. 5a) heterolysis of the carrate
+ COz
(5a)
+[R+ -0zC-SR’I +[R+ -SR’] + COz
(5b)
R-OC-SR’
11
+C o n + [R+ -SR’] +R-SR’ detg.
0 rate
R-0-C-SR‘ I1 0
i
I
4
R-SR‘
1.00
i O ‘ *
Iw’R‘
Figure 2. Log kl for benzhydryl thiocarbonates Ph2CHOCOSR’ vs. U* for R’. p* for S-alkyl and aralkyl thiocarbonates equals +1.54.
decomposition closely parallels that observed in the other two reactions. Taken together with the observed influence of solvent ionizing power, this seems to require that the rate-determining step of the aralkyl thiocarbonate decomposition, like the chlorocarbonate decomposition, involve formation of the aralkyl carbonium ion by heterolysis of the aralkyl-oxygen bond. Variation of Rate with Thioalkyl Group. The influence of the nature of the thioalkyl group, -SR’, on the rate of decomposition of a series of benzhydryl thiocarbonates, PhzCHOCOSR ’, was investigated. The results are summarized in Table IV. 1736
detg.
bonyl-sulfur bond is synchronous with the fission of the R-0 bond. With either mechanism one would expect that the rate of decomposition of ROCOSR’ should increase systematically with an increase in the electronwithdrawing character of R’, since the greater the electron-withdrawing inductive effect of R’ the better it can help to stabilize the negative charge on the anionic fragment of the incipient ion pair. The results in Table IV and Figure 2 are in accord with this expectation. Furthermore, the good correlation between log k , and u* for the various alkyl and aralkyl substituents, despite their varied steric requirements, argues that for the R’ groups in Table IV the steric bulk of R ’ has no significant influence on decomposition rate. Thus the deviation of the data for the S-phenyl ester from the log kl-u* correlation for the remaining esters apparently cannot be ascribed to a steric effect. However, if the mechanism of the decomposition involves heterolysis of the carbonyl-sulfur bond in the rate-determining step, as in eq. 5a, the positive deviation of the S-phenyl ester data is understandable, because the phenyl group, unlike the other substituents in Table IV, can stabilize the R’S- anion not only inductively but also by resonance. (4) The u* values for all the substituents except CH300CCHz- were taken directly from the table given by Taft.6 For CHaOOCCHz- the values for CHaOOC- was divided by 2.8. (5) R. W. Taft, Jr., “Steric Effects in Organic Chemistry,” M . Newman, Ed., John Wiley and Sons, Inc., New York, N. Y . , 1956, Chapter 13, p. 619.
Journal ofthe American ChemicalSociety / 87.8 / April 20, 1965
Convincing evidence for the ability of phenyl to stabilize a mercaptide ion in this manner has already been provided by the results of Kreevoy, et a1.,6 on the acidity of a series of mercaptans. They found an excellent correlation between pKafor R’SH and 6” for R’ for a large number of alkyl and aralkyl mercaptans. However, thiophenol (R’ = phenyl) was 1.6 pK units more acidic than expected from the pKa-a* correlation line for the other mercaptans. Although this enhancement of acidity is less than found for phenol vis-&vis alcohols, it is still significant, and as Kreevoy, et pointed out, it must almost certainly be due to stabilization of the mercaptide ion by delocalization of the negative charge into the ring. The observed deviation of the S-phenyl ester point in Figure 2 thus speaks strongly for a mechanism for the thiocarbonate decomposition involving at least partial heterolysis of the carbonyl-sulfur bond in the rate-determining step. In this connection, however, one should note that a plot of log kl for thiocarbonate decomposition rate vs. pKa of R’SH (Figure 3) shows that the phenyl group does relatively less to enhance the decomposition rate of ROCOSC6Hbthan it does to enhance the acidity of C6HbSH. One possible interpretation of this result is to assume that in the thiocarbonate transition state the breaking of the carbonylsulfur bond lags somewhat behind the cleavage of the aralkyl-oxygen bond, with the result that part of the developing negative charge resides on the oxygens. The transition state for the thiocarbonate decomposition (I) would then be pictured as [R+. . . . 6-OzC - - - 9 -SR’] I The observed solvent and structural effects could also be accomodated by a mechanistic sequence of the form ki
R-0-C-SR‘
I1
J’[R+ -02C-SR’] -% COz + [Rf-SR’] --+ R-SR’ ( 5 ~ ) &n
L
.-I
t YhGHe1
0.4
10
E
pKa fbr R’SH
Figure 3. Log kl for benzhydryl thiocarbonates PhzCHOCOSR’ Y S . Kap for R’SH.
state for the rate-determining step where both R-0 and C-C1 bonds are broken are probably a more accurate representation of the mechanism of the chloro-
carbonate decomposition than is eq. 2b. It would also appear that mechanisms involving at least partial scission of more than one bond in the rate-determining step may be a quite general phenomenon in SNi reactions.
k-1
0
where kz < kF1. Distinction between this mechanism and eq. 5a should be possible through determination of whether or not partial decomposition of carbonyllabeled ROC018SR’ leads to any equilibration of the oxygen label in the recovered undecomposed thiocarbonate. In any event, in contrast to eq. 5b, both eq. 5a and 5c involve fission of the C-S bond in the rate-determining step. The present results thus indicate that, regardless of the specific timing involved, both the R - 0 and C-S bonds are effectively broken by the time the transition state of the rate-determining step of the thiocarbonate decomposition is reached. Naturally this does not necessarily require that a similar situation prevails in the chlorocarbonate decomposition. However, when the present results are taken together with the recent study of Rhoads and Michel,’ wherein those authors found strong evidence for cleavage of more than one bond in the rate-determining step of the decomposition of aralkyl chloroglyoxalates (eq. 6), they certainly suggest that equivalent mechanisms involving a transition (6) M. M. Kreevoy, E. T. Harper, R. E. Duvall, H. S. Wilgus, 111, and L. T.Ditsch, J . Am. Chem. Soc., 82, 4899 (1960). (7) S.J. Rhoads and R. E. Michel, ibid., 85, 5 8 5 (1963).
Experimental Preparation of Thiocarbonates. The following general procedure was employed for the synthesis of the various thiocarbonates. A solution of 0.1 mole of the appropriate alcohol and 0.1 mole of pyridine in 40 ml. of benzene was placed in a three-necked flask which had been fitted with a reflux condenser and drying tube, a dropping funnel, and a stirrer. The solution was cooled in an ice bath and 0.1 mole of the proper chlorothiolformate was added dropwise with stirring. After the addition was complete the mixture was heated to reflux for 2 hr. It was then cooled to room temperature, and 50 ml. of ether was added. The ether-benzene layer was washed three times with water, the aqueous layer being discarded each time. The organic layer was then dried over sodium sulfate, and the solvent was removed under reduced pressure. The crude thiocarbonate was purified by chromatography on alumina, using benzene as eluent, followed by subsequent recrystallization from hexane or hexane-benzene mixtures. Table V summarizes the data on the physical properties, etc., of the various thiocarbonates involved in the present work. Chlorothio2formates. Phenyl, ethyl, and methyl chlorothiolformates were a gift from the Stauffer
Kice, Bartsch, Danklefl, Schwartz
1
Decomposition of Aralkyl Thiocarbonates
1737
Table V. Preparation and Properties of Aralkyl Thiocarbonates -Thiocarbonate, R
ROC0SR'-R'
M.p., "C.
Yield,
87-88 72-73 37-38 45-46 115-118 42-43 50-51 95-96
55 24 22 20 43 26 40 26
CHs
62-63
24
CBH5
59-60
20
C6H5 CHI CHsCHz CeHsCHz (CeH&CH CICHzCHz CHaOzCCHz C6H5 C6H5 p-CICeH4CH
C
Calcd., %---H
75.00 69,74 70.56 75.41 78.99 62.65 64.54 67.70
5.04 5.46 5.92 5.44 5.40 4.93 5.10 4.26
Ci,Hi'CIOzS
61.53
4.48
Ci4Hiz02S
68.85
4.92
--
2
Formula
Analyses -Found, S C 9.99
7
%--H
S
75,08 69,77 70.43 75.51 78.67 62.65 64.65 67.48
5.25 5.44 5.92 5.49 5.27 5.32 4.78 4.36
10.19
61.50
4.48
69.05
5.02
I
C6H5 CeHd&
Yield of pure thiocarbonate.
12.89
Yields of crude thiocarbonate considerably higher.
Chemical Co. They were fractionally distilled before use. Benzyl chlorothiolforrnate was prepared using a modification of Tilles' procedure.8 Phosgene (22.8 g.) was dissolved in 80 ml. of toluene. The mixture was kept at -20-0' and 20.0 g. of benzyl mercaptan was added dropwise. The mixture was stirred for 2 hr. at this temperature, warmed to room temperature, and allowed to stand overnight. Enough aqueous sodium hydroxide was added for the aqueous layer to remain basic to litmus. The toluene layer was then separated, the toluene was removed, and the residue was fractionally distilled, giving 23.5 g. (78 %) of benzyl chlorothiolformate, b.p. 128-129" (12 mm.). Benzhydryl chlorothioljormate was prepared by the same method as benzyl chlorothiolformate. The residue after removal of the toluene was not distilled, but was used directly for the synthesis of the thiocarbonate. An 83% yield of the crude chlorothiolformate was obtained. The benzhydryl mercaptan required for its synthesis was prepared as previously d e ~ c r i b e d . ~P-Chloroethyl chlorothiolformate was synthesized by the procedure of Ringsdorf and Overberger,Io b.p. 57-60' (4-5 mm.). Carbomethoxymethyl Chlorothiolformate. Methyl thioglycolate (23.3 g.) was dissolved in the minimum volume of ethanol and added with stirring to 41.8 g. of lead acetate trihydrate dissolved in 1 : 1 aqueous ethanol. The lead salt of the mercaptan precipitated immediately and was removed by filtration, washed three times with anhydrous ether, and dried. The dried salt (44.3 g.) was added in portions to 22.1 g. of phosgene in 100 ml. of dry toluene. The reaction mixture was kept at 0" for 2 hr. and then left overnight at room temperature. The excess phosgene was removed under reduced pressure, and the crude mixture was filtered to remove lead chloride. After removal of the toluene, the residue gave on distillation 25.3 g. (7 1 %) of carbomethoxymethyl chlorothiolformate, b.p. 63'(1.0-1.2 mm.). Product Studies of Thiocarbonate Decomposition. The reaction flask was the type previously used for study of the thermal decomposition of thiolsulfonates. l (8) H. Tilles, J . A m . Chem. Soc., 81, 714 (1959). (9) J. L. Kice and F.M. Parham, ibid., 82, 6168 (1960). (10) H . Ringsdorf and C. G. Overberger, Makromol. Chem., 44, 418 (1961). (11) J. L. I
